In order to address issues of opacity in production information and difficulties in collecting equipment data in shipbuilding workshops,a digital twin ship manufacturing workshop monitoring system is designed based on the Unity physics platform.The essential steps in building a virtual reality platform are outlined,encompassing the creation of a virtual ship workshop,the development of data transmission methods for multi-source heterogeneous data acquisition,implementation of data-driven methods for achieving virtual-real synchronization,and enhancement of data visualization capabilities.By practically designing a real-time monitoring system for the welding assembly line production process in shipbuilding,the 3D scene reproduction of the welding assembly line and production monitoring of the digital twin workshop in the virtual scene are achieved.The feasibility of constructing this system is validated through the consistency verification between the virtual and physical workshops.
Taking inspiration from the natural gait transition mechanism of quadrupeds, devising a good gait transition strategy is important for quadruped robots to achieve energy-efficient locomotion on various terrains and velocities. While previous studies have recognized that gait patterns linked to velocities impact two key factors, the Cost of Transport (CoT) and the stability of robot locomotion, only a limited number of studies have effectively combined these factors to design a mechanism that ensures both efficiency and stability in quadruped robot locomotion. In this paper, we propose a multi-gait selection and transition strategy to achieve stable and efficient locomotion across different terrains. Our strategy starts by establishing a gait mapping considering both CoT and locomotion stability to guide the gait selection process during locomotion. Then, we achieve gait switching in time by introducing affine transformations for gait parameters and a designed finite state machine to build the switching order. Comprehensive experiments have been conducted on using our strategy with changing terrains and velocities, and the results indicate that our proposed strategy outperforms baseline methods in achieving simultaneous efficiency in locomotion by considering CoT and stability.
In this paper, a Digital Twin (DT) method is proposed for quickly predicting the stress fields of deep-diving spherical shells based on a simulation database. The DT method can completely cover the stress field distribution of the entire pressure-resistant shell and provide stress predictions for different sizes of deep-diving protection shells. The DT model has a three-level structure. The Level-1 DT maps the Finite Element Model to the digital model, which is verified by experiments. The difference between the experimental and numerical results is less than 9.4%. Then, the Level-2 DT deduces the digital model and expands the simulation database samples. The uncalculated stress field results in the simulation database could be obtained by the local Lagrangian interpolation method. Furthermore, the Level-3 DT can rapidly predict the stress field distribution in deep-diving spherical shell digital models. By comparison with different optimization algorithms, the Particle Swarm Optimization algorithm demonstrates strong optimization capabilities and fast convergence, and the error compared with the simulation results is less than 1%. Finally, combined with actual sensor measurements and historical experience data, this method can be used to carry out real-time mapping and synchronous deduction of the service state of the dynamic stress digital twin.
Cooperative vehicle infrastructure system(CVIS)is one of the advanced solutions to enhance intersection vehicle passage safety.Due to the lack of clear specifications and standards regarding the dynamic timing and transition processes of system object state interaction in existing CVIS technologies,ensuring the safety of passage control logic is challenging.This study utilizes formal language to describe the functional logic of CVIS in unsignalized intersections,verifying the safety of system object state interaction and control logic to improve vehicle passage safety at unsignalized intersections.Simulations are conducted for scenarios including single-vehicle non-conflict,dual-vehicle conflict,and multi-vehicle conflict to identify state interactions and enable transition paths.By integrating tools and requirement specification statements for system safety attribute verification,the reliability and safety of control logic are demonstrated,providing a credible basis for developing high-security architecture for CVIS.
The autonomous control of flippers plays an important role in enhancing the intelligent operation of tracked robots within complex environments. While existing methods mainly rely on hand-crafted control models, in this paper, we introduce a novel approach that leverages deep reinforcement learning (DRL) techniques for autonomous flipper control in complex terrains. Specifically, we propose a new DRL network named AT-D3QN, which ensures safe and smooth flipper control for tracked robots. It comprises two modules, a feature extraction and fusion module for extracting and integrating robot and environment state features, and a deep Q-Learning control generation module for incorporating expert knowledge to obtain a smooth and efficient control strategy. To train the network, a novel reward function is proposed, considering both learning efficiency and passing smoothness. A simulation environment is constructed using the Pymunk physics engine for training. We then directly apply the trained model to a more realistic Gazebo simulation for quantitative analysis. The consistently high performance of the proposed approach validates its superiority over manual teleoperation.
For the problem of adaptive motion control of omnidirectional mobile robot under unknown disturbance, in this paper the characteristic modeling and adaptive motion tracking control method of Mecanum-wheels omnidirectional mobile robots are studied. Firstly, the dynamic model of mecanum mobile chassis was established, and the input-decoupling characteristic model was established according to the characteristics of the dynamic model. Then, the Mecanum-wheels omnidirectional mobile robots adaptive motion tracking control method was designed based on characteristic model combined with the golden section control law. Simulation experiments results validate the effectiveness and robustness of the method in presence of uncertainties and unknown external disturbances.
To enhance the realism of garment and human body collision in real-time fabric simulation， this paper proposed an automated human body fitting collision method based on bounding volume and mesh. Firstly， according to the human skeletal structure and garment type， we optimize the skeletal information involved in collision simulation. It is convenient to obtain the feature points and semantically segment the human body. Secondly， the geometric shape of the human body is approximated with capsule colliders and mesh colliders according to the characteristics of skinning animation， which can follow the movement of the model. Lastly， the capsule colliders can realize rough collision detection rapidly and eliminate mesh colliders that are unlikely to detect. It achieves accurate collision based on the mesh collision method and reduces penetration. We use sphere colliders at joints with large deformation. The experiment demonstrates our approach can generate colliders automatically. It improves the fidelity of garment collision simulation and guarantees the efficiency of 3D garment simulation in real-time.
基于现有针对单个风电场的等值建模方法,提出一种适用于风电集群的等值建模方法,利用CloudPSS-XStudio套件完成了风电集群等值建模系统开发,该系统将预想故障选择、等值参数计算和结果分析进行了集成,为含大规模风电集群的电力系统的动态安全分析提供了支撑.以各风电场平均风速以及预想故障作为输入,基于等值模型的迭代仿真得到各个风电机组的分群指标,并完成对风电集群的分群等值建模.以某实际风电集群拓扑接入IEEE 39节点系统作为测试算例,验证了所提方法的正确性.
Simulating the dynamics of fluid flows accurately and efficiently remains a challenging task nowadays, and traditional fluid simulation methods consume large computational resources to obtain accurate results. Deep learning methods have developed rapidly, which makes data-based fluid simulation and generation possible. In this paper, a motion prediction algorithm for long-term fluid simulation is proposed, which is based on a density field with a single frame and a previous velocity field of a sequence. The model focuses on matching the velocity and density fields predicted by the neural network with the simulated data based on the Navier-Stokes equation at a macroscopic level. With the help of fully convolutional U-Net-based autoencoders and LSTM-based time series prediction subnetworks, the model better maintains the visual macroscopic similarity during temporal evolutions and significantly improves computation speed. As a result, the proposed method achieves accurate and rapid long-term motion prediction for the macroscopic distributions of flow field evolution. In addition, the paper demonstrates the effectiveness and efficiency of the proposed algorithm on a series of benchmark tests based on two-dimensional (2D) and three-dimension (3D) simulation data.
航空发动机的研发具有精密度高、跨学科的特点,为减少沟通成本、直观展示发动机结构及半物理仿真机运行状态,应用虚拟现实技术,搭建了一套沉浸式视景仿真系统.通过研究CAD数据轻量化技术、基于物理的实时渲染技术,提出了一种相似对象动态合批的渲染优化方法,有效提高了渲染帧率;提出一种动态视差调整算法,解决了近距离观看立体画面时出现的眩晕问题.该系统实现了沉浸感强、流畅度高、精度高的视景效果,相关技术可应用于发动机沉浸式评审、虚拟培训等场景,为虚拟现实技术应用于制造领域提供技术支撑和参考.
The production logistics mode of manufacturing industry is developing rapidly,on which the modeling and simulation can provide the decision support for the design,analysis and transformation of manufacturing system.A description of the entity elements in intelligent workshop manufacturing system is given according to the classification of"human machine material environment rule".A production and logistics componentized EFSM model is created on the basis of EFSM and componentized modeling ideas.The modeling process for multi-job production in smart shop and the component model instantiation methodology are elaborated.The simulation running through the automatic conversion of EFSM-DEVS model and DEVS engine is completed.The simulation results show that the model established by this method is more in line with the actual situation of the workshop and is more applicable.The idea of component modeling can construct the more scalable software;Modeling and the simulation running of 3D visualization makes the software more intuitive,and the simulation results are consistent with AnyLogic.
To solve unrelated parallel machine scheduling problem(UPMSP)with additional resource and learning effect,a dynamical artificial bee colony(DABC)algorithm is proposed to minimize the makespan.A new representation and decoding process is given and two initial bee swarms are constructed.A swarm evaluation method is applied to dynamically decide employed bee swarms and onlooker bee swarms.Employed bee phase and onlooker bee phase are implemented in different ways to increase exploration ability.The experimental results show that the new strategies of DABC are effective and reasonable,and can obtain results with better convergence,average value and stability,which d has high search performance in solving the considered UPMSP.
Aiming at the offloading and execution of delay-constrained computing tasks for internet of vehicles in edge computing,a task scheduling method based on deep reinforcement learning is proposed.In multi-edge server scenario,a software-defined network-aided internet of vehicles task offloading system is built.On this basis,the task scheduling model of vehicle computation offloading is given.According to the characteristics of task scheduling,a scheduling method based on an improved pointer network is designed.Considering the complexity of task scheduling and computing resource allocation,the deep reinforcement learning algorithm is used to train the pointer network.The vehicle offloading tasks is scheduled by the trained pointer network.The simulation results show that with the same computing resources of edge servers,the proposed method is better than other methods in processing the number of delay-constrained computing tasks,and effectively improves the service capability of the internet of vehicles task offloading system.
To make full use of the massive operation data of automated container terminals and further realize the digital and intelligent transformation of terminals driven by digital twin,a method for digital twin data modeling and effect verification and evaluation of automated container terminals is proposed.The application framework and operation mechanism based on digital twin are studied.Based on the data processing logic of digital twin framework,a method of terminal operation process evolution and dynamic data modeling based on digital twin is proposed.To verify whether the data could meet the effective operation of the digital twin,a quantitative calculation method for the verification and evaluation of terminal digital twin data is given.The research provides method to realize the intelligent operation control and operation data management of automated container terminals based on digital twin.
In order to improve the accuracy of automatic target recognition and promote the effect on substation operation and maintenance,automatic target recognition of substation 3D scene for digital twin is proposed.The automatic target recognition model for the three-dimensional scene of the substation is constructed.The perception module of the model is used to collect the real-time status data of substation,and the communication module is used to transmit the data to digital twin modules.This module,based on the received data information,realizes the deep fusion and panoramic mapping of substation information through the knowledge base constructed by the knowledge map and the virtual and real data processing unit of the service module.Non-uniform rational b-splines(NURBS)surface reconstruction method is used to create a three-dimensional scene model of substation.The forward looking automatic target recognition unit uses the forward looking template matching technology to accurately recognize the single target and multiple targets in the three-dimensional scene model of the substation according to the knowledge base.The experimental results show that the model has excellent 3D modeling effect in both simple and complex scenes.In cloudy and foggy days,the model can still accurately identify all the targets in substation.
Following the large-scale entry of distributed new energy into the network,the uncertainty factor of the distribution network increases significantly,and the difficulty of reactive power optimization scheduling increases accordingly.Traditional optimization solutions have many limitations and shortcomings,and a dynamic reactive power optimization scheme for active distribution networks based on a multi-scenario approach is proposed.The mathematical modeling is carried out separately for the uncertainty of new energy and load,and the multi-scenario method is used to transform the uncertainty problem into a deterministic problem.A mathematical model is constructed on the distribution network side to pursue the integrated optimal value of the expected cost of network loss and reactive power compensation equipment regulation,and the coronavirus herd immunity optimizer is used to solve it.The results show that the optimization scheme obtained from the algorithm can effectively save the distribution network operation cost and reduce the network loss.
Reinforcement learning simulation platform can be an interactive and training environment for reinforcement learning.In order to make the simulation platform compatible with the multi-agent reinforcement learning algorithms and meet the needs of simulation in military field,the similar processes in multi-agent reinforcement learning algorithms are refined and a unified interface is designed to embed and verify different types of deep reinforcement learning algorithms on the simulation platform and to optimize the back-end service of the simulation platform to accelerate the training process of the algorithm model.The experimental results show that,by unifing the interface,the simulation platform can be compatible with many different types of multi-agent reinforcement learning algorithms,and the algorithm training efficiency can be significantly improved after the back-end service framework reconstruction and parameter quantization.
It is difficult to quantitatively calculate and display the real-time traffic situation of expressway ETC system,and there is no simulation system for ETC to optimize the operating situation.A simulation system based on ETC data in proposed,in witch there are three key algorithms.ETC data feature extraction algorithm provides the feature of generating simulation data for the simulation platform.The improved multitask scheduling algorithm has the computing ability of muhitasks in simulation environment.The algorithm of expressway traffic flow control strategy provides the decision index for traffic flow control on the way.The experimental results show that the proposed system can effectively extract the features of ETC data,and is better than other algorithms,which can realize the overall traffic flow control and optimization of simulation road network.
To address the issue of feature loss that occurs during the extraction and transmission of target features in 3D object detection tasks using point cloud data,this study proposes an object detection method based on cross-module attention.This method incorporates a channel attention module and a spatial attention module to enhance the crucial feature information.Through feature transformation,the features from different stages of the attention module are connected to mitigate the loss of features during the extraction and transmission process.To tackle the problem of inadequate detection performance in target detection networks for objects of different scales,a cross-scale feature extraction and fusion method is introduced.This method enhances the network's ability to acquire multilevel features by employing multi-scale feature extraction and fusion techniques.Experimental results demonstrate that the proposed method achieves state-of-the-art performance while maintaining a real-time inference speed of 33 Hz.
Due to the complex task,tight coupling,strict timing,and a large amount of interchange data,the technical threshold of automated testing of bus communication equipment software is high,and the implementation is difficult.The ideas of data-driven testing and keyword-driven testing are introduced,and a data simulation testing framework is proposed.Configuration rules are formulated and implemented in the framework.Testers can simulate peripheral data for complex process equipment software and implement automated testing by only focusing on the task analysis,and configuring interchange data and keywords.There is no need to develop test scripts,which reduces the technical threshold of automated testing.Application results show that the framework can perform more task scenarios than manual and physical testing.It also can find the deep software defects related to the task processes.
